诉讼称xAI使用儿童性虐待材料训练Grok模型

一名自称Jane Doe的受害人周三提起诉讼，指控xAI公司在训练Grok AI模型时使用了儿童性虐待材料（CSAM）1。根据诉讼，Doe在2000年代初期遭受性虐待，其被制作成的CSAM图像被用于模型训练过程中1。

加拿大儿童保护中心（CCCP）在xAI平台上发现了描绘Doe的AI生成虐待内容1。此外，在线论坛中出现了犯罪分子讨论为Doe和其他受害者创建AI生成CSAM的消息1。Doe通过美国司法部受害人通知系统收到了相关警报1。

xAI faces serious allegations that it utilized child sexual abuse material (CSAM) in training its Grok artificial intelligence models, according to a lawsuit filed Wednesday.1 A plaintiff identifying herself as Jane Doe alleges that images of her abuse from the early 2000s were incorporated into the model's training data, and that AI-generated abusive content depicting her subsequently appeared on xAI's platform.1 The Canadian Centre for Child Protection discovered the AI-generated CSAM images on xAI, and Doe received notifications about the case through the U.S. Department of Justice's victim notification system.1

The lawsuit also reveals that criminals discussed creating AI-generated CSAM involving Doe and other victims across online forums.1 These allegations raise significant concerns about content moderation practices and the sources of training data used by major AI developers.